DISSIMILATION
"A-A", but more often "A-B".
In sounds, syllables, onsets, words, phrases, clauses, sentences, chapters, novels.
We linguists, like many other scholars, are fond of generalizations, we prefer descriptions (or rules) that work in a lot of cases, rather than rules with a narrow scope.
Talking of non-repetition or dissimilation, I believe that the following is one of the most general descriptions possible.
SOUNDS. When an utterance or text (which virtually is the same) starts, the first element (sound or letter) can be 'a'. What is the probability that the next element be 'a' ? Very low. First, in general, repetitions of exactly the same element are rare, secondly most of the time, after a vowel comes a consonant (in English vowels are most simply written a,e,i,o,u). Viceversa, if we start with 'b' (consonant), we should expect the next element to be a vowel (a,e,i,o,u). Between the two categories vowel and consonant there is a preference or, in other words, vowels "repulse" vowels and consonants "repulse" consonants, and we call that non-repetition. (This alternation of vowels and consonants in a spoken or written string contains a germ of the definition of the two categories.)
The same happens inside the two categories: after a 'd' we don't expect another 'd', but another consonant - or a vowel. Sequences like aadi or addi or adii are rare (in English). To be sure, in the written language you find almost everything, e.g. 'oo' for /u/, 'ee' for /i/, but that is only graphic, in English pronunciation /aa, dd, ii/ are extremely rare. This at the phonetical/phonemic level.
SYLLABLES the same. A word with the structure C¹V¹C¹V¹ (e.g. mama), with two indentical syllables, is rarely found, in the lexicon (such a word can be frequently used in young families, e.g. daddy), most words have other structures, like C¹V¹C²V¹ (mini, in ordinary orthography), C¹V¹C¹V² (mommy), C¹V¹C²V² (many), and, of course, numerous other ones: C¹V¹C²C³V² (monkey), C¹V¹C²V²C³ (moppet), C¹V¹C²C³V²C° (manhood), and so on, ever more complicated.
The BEGINNINGS OF WORDS (or the onsets of syllables) should vary: Normally people don't say "whiderrwarrd he wendeþþ" (Ormulum) or "And makes as healthful music. It is not madness" (Hamlet III,4), if they are able to avoid it. If someone says: "Our eight oldest animals enjoy eating eggs every other hour", he/she will hear that it sounds like a poem, or a tongue twister. Here, naturally, many persons will say that this function is conscious, people think about it, when they avoid certain repetitions. That is true, but it also reveals that people have an intuition about the phonetic non-repetitive shape of their language. In reduplication, contrarywise, you find repetition applied deliberately.
WORDS: You can use repetition to emphasize something, or for some other reason - because in all the other cases you don't repeat the words. If you do repeat without intention, it means you are stuttering, or nervous or frightened. Otherwise you just say the word once. Of course many words are employed repeatedly in everyday speech, because they belong to the group of small words that are needed in almost any phrase or clause, like 'the', 'a', 'be', 'as', 'will' etc. A few of them have synonyms, but most of them have not, so you cannot avoid to use them often in your text, you cannot escape repetition. Words like 'that' and 'which' can in many cases substitute each other, 'you can..' may be replaced by 'we can..' or 'you might'; the word 'substitute' can be used instead of 'replace'. In the grammar of many languages you can apply repetition on purpose, to indicate plural of a noun (e.g. Javanese rumah-rumah 'houses') or a high degree of an adjective or adverb.
PHRASES, CLAUSES: Sometimes you repeat an expression to make sure that the listener has heard it. Apart from that, just as words, the repeated phrase is odd, when used there must be a reason, the speeker want to add stress to the expression, or there is another, more or less stylistic cause. Otherwise you don't repeat your phrase.
SENTENCES: The higher in the hierarchy of linguistic elements, the less probable is repetition. If you see a sentence repeated immediately in your newspaper, you intuitively think it's a misprint. The same goes for a CHAPTER in a novel, or a whole NOVEL: If you find the same novel in another presentation, it must be a reprint, a paperback edition, or a fraudulent reuse of the same story.
All these shortly described processes, phonological, morphologic, derivational and syntactic, are non-repetitive, no matter how they show themselves in details.